mlsanityData to 7 Oct 2026

Mind Your Steps: A General Learning Framework for Accurate Humanoid Foothold Tracking

arXiv 2606.08253 ↗First posted 6 Jun 2026cs.RO cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Enabling humanoid robots to operate in complex, dynamic environments remains a critical challenge, fundamentally limited by the ability to navigate robustly, safely, and accurately. While reinforcement learning with velocity-commanded policies has achieved remarkable robustness in humanoid locomotion, this approach lacks explicit control of the foothold placement, leading to unsafe behavior, such as stepping onto human feet, or imprecise navigation, hindering the following manipulation task. Conversely, explicit foothold-tracking policies offer a promising alternative by directly being commanded with target foot poses. However, existing approaches are often limited by unrealistic state assumptions, compromising real-world deployment, or they are part of staged pipelines, making them tied to specific downstream tasks. In this work, we introduce a novel, lightweight framework for training general-purpose 3D foothold-tracking policies. By dynamically providing footstep support through a goal sampler, this method enables the learned policy to be agnostic to specific terrains. Our new target representation effectively mitigates challenges arising in the real world, such as noisy and inaccurate pose estimation and foot contact estimation. Designed for direct real-world transfer, our policy acts as a standalone low-level controller that can be seamlessly paired with various high-level foothold generators. We demonstrate the effectiveness of our framework through extensive experiments in simulation and in the real world. By coupling our policy with different upstream planners, we achieve natural and accurate locomotion in challenging settings, paving the way for loco-manipulation tasks in complex environments.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 39 of its 53 references that are in the map. First posted on arXiv 6 Jun 2026.

Builds on

39 of 53 references.

References · 39

  1. 2017Proximal Policy Optimization Algorithmscited by 14,441
  2. 2020Learning Quadrupedal Locomotion over Challenging Terraincited by 494
  3. 2021RMA: Rapid Motor Adaptation for Legged Robotscited by 430
  4. 2020Curriculum Learning for Reinforcement Learning Domains: A Framework and Surveycited by 350
  5. 2022Learning robust perceptive locomotion for quadrupedal robots in the wildcited by 299
  6. 2018Residual Reinforcement Learning for Robot Controlcited by 242
  7. 2017Asymmetric Actor Critic for Image-Based Robot Learningcited by 202
  8. 2018Residual Policy Learningcited by 146
  9. 2022Legged Locomotion in Challenging Terrains using Egocentric Visioncited by 131
  10. 2022Rapid Locomotion via Reinforcement Learningcited by 130
  11. 2023Extreme Parkour with Legged Robotscited by 120
  12. 2023Robot Parkour Learningcited by 115
27 more
  1. 2020Sim-to-Real Learning of All Common Bipedal Gaits via Periodic Reward Compositioncited by 71
  2. 2024Agile But Safe: Learning Collision-Free High-Speed Legged Locomotioncited by 53
  3. 2020RLOC: Terrain-Aware Legged Locomotion using Reinforcement Learning and Optimal Controlcited by 46
  4. 2019DeepGait: Planning and Control of Quadrupedal Gaits using Deep Reinforcement Learningcited by 45
  5. 2020ALLSTEPS: Curriculum-driven Learning of Stepping Stone Skillscited by 44
  6. 2023Neural Volumetric Memory for Visual Locomotion Controlcited by 40
  7. 2023DreamWaQ: Learning Robust Quadrupedal Locomotion With Implicit Terrain Imagination via Deep Reinforcement Learningcited by 37
  8. 2023DTC: Deep Tracking Controlcited by 37
  9. 2025BeamDojo: Learning Agile Humanoid Locomotion on Sparse Footholdscited by 32
  10. 2023LocoMuJoCo: A Comprehensive Imitation Learning Benchmark for Locomotioncited by 31
  11. 2023Hybrid Internal Model: Learning Agile Legged Locomotion with Simulated Robot Responsecited by 31
  12. 2022Perceptive Locomotion through Nonlinear Model Predictive Controlcited by 27
  13. 2025Attention-Based Map Encoding for Learning Generalized Legged Locomotioncited by 27
  14. 2021GLiDE: Generalizable Quadrupedal Locomotion in Diverse Environments with a Centroidal Modelcited by 23
  15. 2023Learning Agile Locomotion on Risky Terrainscited by 15
  16. 2020Motion Planning for Quadrupedal Locomotion: Coupled Planning, Terrain Mapping and Whole-Body Controlcited by 10
  17. 2021Receding-Horizon Perceptive Trajectory Optimization for Dynamic Legged Locomotion with Learned Initializationcited by 10
  18. 2022TAMOLS: Terrain-Aware Motion Optimization for Legged Systemscited by 10
  19. 2025Gallant: Voxel Grid-based Humanoid Locomotion and Local-navigation across 3D Constrained Terrainscited by 10
  20. 2019Footstep Planning for Autonomous Walking Over Rough Terraincited by 8
  21. 2022Sim-to-Real Learning of Footstep-Constrained Bipedal Dynamic Walkingcited by 8
  22. 2022Learning Bipedal Walking On Planned Footsteps For Humanoid Robotscited by 5
  23. 2024Walking with Terrain Reconstruction: Learning to Traverse Risky Sparse Footholdscited by 4
  24. 2023Perceptive Locomotion through Whole-Body MPC and Optimal Region Selectioncited by 3
  25. 2024Integrating Model-Based Footstep Planning with Model-Free Reinforcement Learning for Dynamic Legged Locomotioncited by 2
  26. 2023SafeSteps: Learning Safer Footstep Planning Policies for Legged Robots via Model-Based Priorscited by 1
  27. 2025Physics-Guided Residual Reinforcement Learning for Humanoid Narrow-Path Traversalcited by 1

Built on by

Not cited yet.